AMD Ryzen 9 9955HX3D vs Intel Core Ultra 9 290HX Plus Comparison

AMD Ryzen 9 9955HX3D and Intel Core Ultra 9 290HX Plus are two flagship mobile processors with contrasting designs. The AMD part uses a 16-core, 32-thread Zen 5 configuration with a massive 128 MB L3 cache, while Intel counters with a 24-core, 24-thread Arrow Lake-HX Refresh design on a 3 nm node. Benchmark results show a near-even split, with AMD taking 8 wins and Intel taking 7 in the head-to-head tests. The data indicates a clear trade-off: AMD dominates integer-heavy and physics workloads, while Intel leads in floating-point math and single-threaded tasks.

FAQ

Q: Which processor has more cores?

A: The Intel Core Ultra 9 290HX Plus has 24 cores, while the AMD Ryzen 9 9955HX3D has 16 cores. However, the AMD chip supports 32 threads via simultaneous multithreading, whereas the Intel chip has 24 threads, matching its core count.

Q: What is the L3 cache difference?

A: The AMD Ryzen 9 9955HX3D has 128 MB of L3 cache, while the Intel Core Ultra 9 290HX Plus has 36 MB of shared L3 cache. This is a significant difference that likely contributes to AMD's strong performance in data compression and integer math.

Q: Which processor has a higher boost clock?

A: The Intel Core Ultra 9 290HX Plus boosts to 5.50 GHz, slightly higher than the AMD Ryzen 9 9955HX3D's 5.40 GHz boost clock. The Intel chip also has a higher base clock at 2.70 GHz versus 2.50 GHz.

Q: How do they compare in overall benchmark percentile?

A: The AMD Ryzen 9 9955HX3D ranks in the 96th percentile of all CPUs, while the Intel Core Ultra 9 290HX Plus ranks in the 95th percentile. Both are near the top of the database.

Q: What is the memory bandwidth for each?

A: The Intel Core Ultra 9 290HX Plus has a memory bandwidth of 102.4 GB/s, while the AMD Ryzen 9 9955HX3D has 89.6 GB/s. Both support dual-channel DDR5 memory.

Q: Which CPU has a smaller manufacturing process?

A: The Intel Core Ultra 9 290HX Plus uses a 3 nm process node from TSMC, while the AMD Ryzen 9 9955HX3D uses TSMC's 4 nm node. The smaller node generally allows for better power efficiency.

Architecture Differences

The two processors represent fundamentally different design philosophies. AMD's Ryzen 9 9955HX3D uses the Zen 5 architecture on the Fire Range codename, built on a 4 nm TSMC process. It packs 16,630 million transistors across a dual-die setup with each die measuring 70.6 mm². The chip features 16 cores with 32 threads, enabled by simultaneous multithreading. Each core has 80 KB of L1 cache and 1 MB of L2 cache, while the shared L3 cache is a massive 128 MB, characteristic of the 3D V-Cache design.

Intel's Core Ultra 9 290HX Plus uses the Arrow Lake-HX Refresh codename under the Core Ultra Series 2 branding. It is built on a smaller 3 nm TSMC process and integrates 17,800 million transistors on a single 243 mm² die. The CPU has 24 cores but only 24 threads, indicating it lacks hyperthreading. Each core gets 192 KB of L1 cache and 3 MB of L2 cache, while the shared L3 cache is 36 MB. This is a notably smaller cache pool than AMD's offering.

The process node difference is significant: 3 nm versus 4 nm. Both processors use TSMC as the foundry. The Intel chip has a higher base clock (2.70 GHz) and boost clock (5.50 GHz) compared to AMD's 2.50 GHz and 5.40 GHz. Both are unlocked for overclocking, and both support ECC memory.

PCIe support differs as well. The AMD processor provides 28 PCIe Gen 5 lanes from the CPU, while Intel offers 20 PCIe Gen 5 lanes. Integrated graphics also differ: AMD uses the Radeon 610M, while Intel uses Arc Xe-LPG Graphics with 64 execution units.

Head-to-Head Benchmarks

The benchmark data reveals distinct strengths. AMD's Ryzen 9 9955HX3D wins 8 of the 15 head-to-head tests, while Intel takes 7. The margins tell the story.

AMD's largest victory comes in PassMark integer math, scoring 222503 against Intel's 164839, a 35% advantage. This is a substantial gap for a workload that stresses arithmetic operations. Similarly, AMD wins PassMark physics with a 38.4% margin (4687 versus 3387). Data compression also favors AMD heavily: 785811 versus 658724, a 19.3% lead. Extended instructions show AMD ahead by 22.5% (62813 versus 51290).

Other AMD wins include PassMark multithread (62674 versus 59439, 5.4% ahead), random string sorting (83497 versus 80327, 3.9% ahead), and find prime numbers (525 versus 519, a slim 1.2% lead). In Cinebench R15 multicore, AMD edges out Intel 6042.5 versus 5981, a marginal 1% difference.

Intel counters with its own decisive wins. PassMark floating-point math shows Intel at 201773 versus AMD's 144122, a 28.6% advantage. Data encryption also favors Intel: 50008 versus 39446, a 21.1% lead. Single-threaded performance is clearly Intel's territory, with PassMark single thread at 4951 versus 4511, an 8.9% difference. Cinebench R23 single core shows Intel ahead by 8.3% (2356 versus 2159.5), and Cinebench R15 single core gives Intel a 2.4% edge (340 versus 332).

The Cinebench R23 multicore result is interesting: Intel wins 39684 versus 38161.5, a 3.8% margin, despite AMD's strong showing in other multithreaded tests. This suggests the 24-core Intel design excels in rendering workloads that scale well across physical cores.

The Verdict

The data presents a clear choice based on workload priorities. The AMD Ryzen 9 9955HX3D is the stronger option for integer math, physics simulation, and data compression tasks. Its 35% lead in integer math and 38.4% lead in physics are decisive. The 128 MB L3 cache appears to provide a significant advantage in these areas.

The Intel Core Ultra 9 290HX Plus is the better pick for floating-point math, encryption, and single-threaded performance. Its 28.6% lead in floating-point math and 8.9% lead in single-threaded PassMark tests show clear superiority in these domains. Intel also wins in Cinebench R23 multicore, suggesting an edge in rendering workloads.

For general-purpose use, the average benchmark scores differ notably: AMD sits at 97453 while Intel is at 79574. However, the head-to-head record is nearly even at 8-7, meaning the average score alone does not tell the full story. Users who prioritize integer-heavy workloads like database operations or physics calculations should favor AMD. Those who need maximum floating-point throughput or single-thread responsiveness should lean Intel.

Specification Differences

| Specification | AMD Ryzen 9 9955HX3D | Intel Core Ultra 9 290HX Plus |

|---|---|---|

| Cores | 16 | 24 |

| Threads | 32 | 24 |

| Base Clock | 2.50 GHz | 2.70 GHz |

| Boost Clock | 5.40 GHz | 5.50 GHz |

| Process Node | 4 nm | 3 nm |

| Transistors | 16,630 million | 17,800 million |

| Die Size | 2x 70.6 mm² | 243 mm² |

| L1 Cache | 80 KB (per core) | 192 KB (per core) |

| L2 Cache | 1 MB (per core) | 3 MB (per core) |

| L3 Cache | 128 MB | 36 MB (shared) |

| Memory Bandwidth | 89.6 GB/s | 102.4 GB/s |

| PCIe Lanes | Gen 5, 28 lanes | Gen 5, 20 lanes |

| Integrated Graphics | Radeon 610M | Arc Xe-LPG Graphics 64EU |

| Socket | AMD Socket FL1 | Intel BGA 2114 |

| Release Date | 2025-01-05 | 2026-03-16 |

Where Each One Wins

The AMD Ryzen 9 9955HX3D wins in data compression, extended instructions, integer math, multithreaded PassMark, physics, prime number finding, random string sorting, and Cinebench R15 multicore. These are workloads that benefit from large cache capacity and high thread counts. The 128 MB L3 cache clearly helps with data compression and integer operations, where the 19.3% and 35% leads respectively are substantial. Physics simulation also responds well to AMD's architecture, with a 38.4% advantage.

The Intel Core Ultra 9 290HX Plus wins in data encryption, floating-point math, single-threaded tests, Cinebench R15 single core, Cinebench R23 single core, and Cinebench R23 multicore. The floating-point math result is particularly notable, with Intel leading by 28.6%. Encryption performance shows a 21.1% advantage, and single-threaded workloads consistently favor Intel by nearly 9%. The Cinebench R23 multicore win, despite AMD's multithreaded PassMark advantage, suggests Intel's physical cores handle rendering tasks efficiently.

The practical split: choose AMD for simulation, compression, and integer-heavy computational work. Choose Intel for floating-point calculations, encryption, and applications that depend on single-core speed.